Training of deep residual networks with stochastic MG/OPT

Code

• github.com/eulerinstitute/mgopt_icml21
  OfficialIn paperpytorch★ 0

Abstract

We train deep residual networks with a stochastic variant of the nonlinear multigrid method MG/OPT. To build the multilevel hierarchy, we use the dynamical systems viewpoint specific to residual networks. We report significant speed-ups and additional robustness for training MNIST on deep residual networks. Our numerical experiments also indicate that multilevel training can be used as a pruning technique, as many of the auxiliary networks have accuracies comparable to the original network.

Reproductions